Evidence Against Algorithmic Vote Flipping (no fraud)

[da32130]

This thread is created to bring this to grassroots central with new Parts added. Since the other pro flipping posters have created a new thread.

------------------------------------

post without data at bottom

Part 5: How to prove flipping

There has been some confusion about what our demographic graphs mean. And why we feel they are so damaging to the fipping case. I will provide an alternate analysis that hopefully will clear up that confusion. Data is located at bottom

SUmmary: no fraud

The demographic data we have eliminates the vote flipping in the areas it should. We have no reason to believe the same won't be true in other areas when the relevant demographic data is available.

How to prove fraud

In order to show there exists vote flipping you need:
1) total precinct vote to be meaningful on voting results (can be shown in graph or regression form)
2) it needs to remain meaningful even adjusting for demographics (demographics predict total precinct vote, so we would expect total precinct vote to predict voting results in candidates where demographics matter, so unadjusted data doesn't say much)

What do current results tell us

At the county level in SC (nice demographic data) and NH (just indie/dem turnout and libertarian turnout to explain Paul,Huntsman,Romney) step 2) eliminates the value from total precinct vote. So no fraud evidence at the county level.

At the precinct level in SC we have no demographic data to explain Romney and Gingrich. So step 2) can't even take place.

At the precinct level in NH and VA we have indie/dem turnout and libertarian turnout to explain Paul,Huntsman,Romney. This should capture a lot, but isn't a complete demographic analysis. What it misses is exactly what is needed in SC: Income,Age,etc info to separate highly republican areas from eachother. These areas have the greatest turnout and are typically located at the far right in graphs. Based on exit polling we would expect Romney to due unusually well in the highest turnout of the republican areas because they have wealthy older voters who are likely to vote.

How to adjust total precinct vote.

regression:
1) regress demographics and total precinct vote on Romney%, etc

graph:
1) regress demographics on total precinct vote
2) take total precinct vote and subtract the regression result to get the errors
3) do the standard graph but replace total precinct vote with the errors from above

If romney is correlated with the errors it means 1) vote flipping or 2) missing demographics

What are the results for NH and VBC?

Given the insufficient demographic data it isn't a surprise that regression analysis still gives total precinct vote some value. Below we have graphs that show what that value is.

VBC
standard:
http://img820.imageshack.us/img820/2003/vbctv.jpg

total precinct vote excluding impact from available demographic factors (in theory even more would make this chart flatter)
http://img826.imageshack.us/img826/5892/vbcexdem.jpg

NH
standard:
http://img832.imageshack.us/img832/9431/nhtv.jpg

total precinct vote excluding impact from available demographic factors (in theory even more would make this chart flatter)
http://img62.imageshack.us/img62/7772/nhexdem.jpg

So essentially the limited demographic data we have eliminates the value in 85-90% of precincts and in the exact areas we would expect.

Conclusion: no fraud

The demographic data we have eliminates the vote flipping in the areas it should. We have no reason to believe the same won't be true in other areas when the relevant demographic data is available.

------------- review of what data tells us right now:

At the county level in SC (nice demographic data) and NH (just indie/dem turnout and libertarian turnout to explain Paul,Huntsman,Romney) step 2) eliminates the value from total precinct vote. So no fraud evidence at the county level.

At the precinct level in SC we have no demographic data to explain Romney and Gingrich. So step 2) can't even take place.

At the precinct level in NH and VA we have indie/dem turnout and libertarian turnout to explain Paul,Huntsman,Romney. This should capture a lot, but isn't a complete demographic analysis. What it misses is exactly what is needed in SC: Income,Age,etc info to separate highly republican areas from eachother. These areas have the greatest turnout and are typically located at the far right in graphs. Based on exit polling we would expect Romney to due unusually well in the highest turnout of the republican areas because they have wealthy older voters who are likely to vote.

 

[da32130]

Demographic Link - Part 4 - South Carolina 2012

background (includes links to parts 1 and 2)
http://www.ronpaulforums.com/showth...orithmic-vote-flipping.&p=4274995#post4274995

part 3 - Virgina Beach City, VA 2012
http://www.ronpaulforums.com/showth...orithmic-vote-flipping.&p=4279027#post4279027

background on part 4

This analysis is in reply to requests to explain Gingrich's 2012 South Carolina numbers (and should apply to Huckabee's 2008 numbers as well). It is to show that demographics can explain more than just the Paul, Huntsman, libertarians, and Obama relationships in our prior parts.

The hypothesis for this analysis was also alluded to in the initial background link above.


Summary: no fraud

The demographics that are driving turnout are also driving Newts fall as turnout rises.

This shows that demographics can explain not only Paul and Huntsman, but can also explain other candidates.

We believe there is no reason to expect they can't explain other states as well.

To prove fraud one must adjust for these demographics. A downwardly or upwardly sloped graph isn't proof (The initial vote flipper theory cited the downward slope as the primary evidence for fraud). Only if it contradicts the demographics is there a potential for fraud.

Standard Chart

standard analysis (using county data because demographic data is county data)
http://img706.imageshack.us/img706/6950/totalvote.jpg

Regression Analysis

Unlike before I'm coming with the regressions 1st.

When we regressed the available demographic data turnout on we found:

O08P -4.20636
WHITE -0.02383 (highly correlated with Obama, without Obama becomes valuable)
INCOME 4.05716
popSQMILE 3.64639 (
OVER65% 4.53922
FEMALE% -2.13215

Over 2 or less than -2 means it is valuable.

The above is saying that Repub turnout is lower in Obama, poor, rural, younger, female areas.

Now lets look at exit polling:
http://www.cnn.com/election/2012/primaries/epolls/sc

Newt does slightly better with Men than Romney (positive correlation with turnout)
Newt does significantly better with younger voters than Romney (negative correlation with turnout)
Newt does worse with indies than Romney (positive correlation with turnout-based on Obama)
Newt does much better with lower income voters than Romney (negative correlation with turnout)
Newt does much better as population becomes rural vs urban than Romney (negative correlatio with turnout)

So there is a bit of a mixed bag. 3 negatives and 2 positives. However, all positives and negatives aren't created equal. Below we give graphs of each factor. We then give a weighted graph based on the t-stats above. The weighted graph is a demographic based demonstration of what causes turnout.







http://img210.imageshack.us/img210/6144/sqmile.jpg

(notice this graph is actually the opposite of what you would expect for Gingrich, this is due to Paul and Santorum being even more popular in the younger demo than Gingrich)
http://img193.imageshack.us/img193/2715/over65percent.jpg

http://img715.imageshack.us/img715/2664/obamag.jpg

http://img109.imageshack.us/img109/6295/incomer.jpg

http://img4.imageshack.us/img4/8123/femaled.jpg

-------------
This is the weighted graph:

http://img100.imageshack.us/img100/6448/weighted.jpg

What that weighted graph shows is that the demographics that are driving turnout are also driving Newts fall as turnout rises.

There is no need for a vote flipping explanation.

Conclusion: no fraud

The demographics that are driving turnout are also driving Newts fall as turnout rises.

This shows that demographics can explain not only Paul and Huntsman, but can also explain other candidates.

We believe there is no reason to expect they can't explain other states as well.

To prove fraud one must adjust for these demographics. A downwardly or upwardly sloped graph isn't proof. Only if it contradicts the demographics is there a potential for fraud.

 

[da32130]

Demographic Link - Part 3 - Virginia Beach City (VBC), VA 2012

background (includes links to parts 1 and 2)
http://www.ronpaulforums.com/showth...orithmic-vote-flipping.&p=4274995#post4274995

background on part 3

This analysis is in reply to affa's analysis of VBC and his response about my Part 1 on Arlington, VA.

Graphs

standard analysis (ordered by increasing precinct total vote)
http://img405.imageshack.us/img405/6950/totalvote.jpg

obama analysis (ordered by decreasing Obama% vs McCain in 2008)
http://img263.imageshack.us/img263/8600/obamary.jpg

libertarian analysis - includes Barr (libertarian candidate) and also Baldwin (who Paul endorsed) (ordered by decreasing libertarian % vs McCain in 2008)
http://img337.imageshack.us/img337/7465/barrbaldwin.jpg

The t-stat for Obama on total vote is -7 (highly negatively correlated, tstats less than -2 or above 2 are usually meaningful). T-stat for libertarians on total vote is -4 (again highly negatively correlated)

Weighing each by their t-stat (obama 7 weight, libertarians 4 weight) the analysis is done again:
http://img20.imageshack.us/img20/5920/obamabarrbald.jpg

In essense the demographics explain virtually everything. In fact, explain more than looking at total vote based on the smoothness of the last curve.

What does the above look like at the precinct level. I'll give two extreme examples.

The top weighted area for Paul:

Precinct 93 (Newton)

Obama 1092 (Obama crushes McCain)
McCain 116
Barr+Baldwin 5 (almost 5% of the McCain vote, way above normal)

and this translates into

Paul 22 (an absolute drubbing of Romney)
Romney 1

Total Votes 23 (2nd lowest)

The bottom weighted area for Paul:

Precinct 10 (Great Neck)

Obama 647 (Obama gets crushed by McCain)
McCain 1563
Barr+Baldwin 7 (roughly .5% of McCain, 1/10 of where it was in Paul's best area)

and this translates into

Paul 98 (Yes, almost Paul's worst performance)
Romney 238

total votes 336 (the 4th highest vote total)

Everything in between is just a noisy variation on those extremes.

here is the data, look for yourself:
primary 2012
https://www.voterinfo.sbe.virginia....10_A5DCD6FA-6694-4931-BBDB-D87E4356EC47.shtml

general 2008
https://www.voterinfo.sbe.virginia....10_89BE12EC-7BBF-479C-935A-9B8C51DD3524.shtml

What about the late surge for Romney and decline for Paul in the total vote graph?

Notice how before this happens the data is flat. Why can't the surge just be a catch up for the flat period. It is a return to trend.

How about a demographic reason for the extreme shift? Here are the demographics for the 4 precincts right before the shift:

Obama 0.418421053
Libertarian 0.012242518

Here are the 4 right after the shift:
Obama 0.413868117
Libertarian 0.00691085

What happened? Obama was slightly less. But you had a libertarian collapse. So you would expect a surge from Romney.

Just to make this more clear. Here are the average and median libertarian percentage accross all precincts:
average 0.011127577
median 0.010805226

The the split marks where we went from above average to well below average libertarian areas. The declining Obama percentage only made it worse.

What about Arlington?

In Paul's best 5 areas here are the stats:

Paul 0.480692111
Obama 0.715668332
Libertarians 0.026650163
2012 turnout 117.6

In Paul's worst 5 areas:
Paul 0.234412571
Obama 0.603117985
Libertarian 0.015883676
2012 Turnout 216

Exactly as you would expect based on what I've told you.

Flaws in your Arlington analysis

Your whole analysis is dealing with absolute vote counts. Liberty's and RonRules graphs have always been about Paul's % declining, not absolute vote numbers declining.

The reason why Paul does best in Obama areas isn't because Paul's absolute number is higher, but because Romney's number plummets even more. So Paul's percentage is higher. This is right in line with Obama being inversely related to turnout

Conclusion: No Fraud. (affa, I don't mind if you do this either. I think it gives a good quick read for people.)

In both Arlington and Virginia Beach Paul's numbers are explained by demographics. Total precinct vote is actually a worse measure than demographics. Why? because the demographics are the driver. The total vote numbers are just correlated with the demographics.

 

[da32130]

Demographic Links - Part 2
New Hampshire

Standard Analysis
http://i.imgur.com/NO2Al.jpg

Obama % (hurts huntsman)
http://i.imgur.com/we76k.jpg

Paul not hurt using Obama, maybe due to huntsman. An alternative the relative performance of Barr (libertarians) in 2008 vs McCain. While in a general election libertarians are only a few percent the data should pick up demographic information that would enhance Paul's turnout. The following graph includes that in the standard analysis.

Libertarian vs McCain full
http://i.imgur.com/PEWMh.jpg

There is a lot of noise in the first half. Using just the second half brings out some clear trends. Barr and McCain (normalized for vote %) exhibit the same behavior as Paul and Romney.

Lib (barr) vs McCain last half
http://i.imgur.com/AJpDd.jpg

Combining both Obama and Libertarian turnout explains the 2nd half fall of Paul and Huntsman. The 1st half is muddier but there are no clear indications of fraud.

Conclusion NH: no fraud

I welcome feedback. The conclusion is just based on this analysis. Although, I believe it is likely future analysis will back this up in other states, etc.

 

[da32130]

Demographic Link Found

The first graph is the standard analysis with downward sloping ending vote totals for Paul:
http://i.imgur.com/shYwJ.jpg

This second graph is the cumulative vote total for Paul but starting with the precincts with the highest Obama percentage vs McCain in 2008.
http://i.imgur.com/7Cw11.jpg

The reason why they are so close is that Obama 2008 % is negatively correlated with republican turnout in VA for 2012.

The t-stat on Obama 2008 % as a forecast for total Republican votes in VA for 2012 is -3.51.

The t-stat for Paul 2012 % is -3.72.

If you look at the turnout % of the primary vs the general the Obama 2008 % t-stat is -8.9, while for Paul 2012 % it is -3.47.

So when Obama does well in an area you can imply that Republican turnout is lower and Paul should better in that area. Creating the anomaly of the downward slope to Paul vote % as turnout is higher.

I would encourage others to replicate my analysis on the most damning evidence of fraud. Alternatively, I may be willing to do so as well.

Conclusion: No fraud. Just low turnout in Obama and Paul friendly areas.

 

[da32130]

umm... I just see "no fraud" bolded everywhere, but no real explanation... I'm also pretty sure we weren't exactly looking for FRAUD, since fraud is a very broad word for what is possible to view through statistics. I'm sorry I couldn't go through every post of the last 10 pages, can we just have a clean discussion about each side of the argument?

soulcyon,

I've tried to give a beginner's summary of what is going on.

Background

As precinct (one step smaller than a county, each county has many precincts, each state has many counties) vote totals rises Paul's cumulative percentage falls in many places. Essentially he does better in precincts where there are less total votes. While getting crushed where precinct vote totals are the largest.

The candidate who typically gains from this is Romney. Whos cumulative percentage rises as precincts get larger.

This could be fraud if votes are flipped from Paul to Romney once turnout is high enough.

Why it matters

If there isn't a demographic link for the total vote percentage different then it would be a case were in states like NH, Iowa, SC, and VA Paul would have lost 5,10,15% + relative to Romney. In some cases deciding the election.

Demographic Link

I think it has been agreed upon that if there is no demographic link to explain varying total votes in precincts then fraud is highly likely.

Assuming precincts are all roughly the same size an urban/rural divide doesn't explain the differing percentages (although this could be looked into more).

VA

What I have found is that in places like VA the total vote in a Republican Primary is inversely proporational to how well Obama did in a precinct during the 2008 general election vs McCain. My conclusion is that the best precincts for Obama have the fewest Republicans. So in a Republican Primary those areas will have the smallest vote total.

It well known through exit polling Paul does best with Independent voters and Dems relative to the other candidate but struggles with the Republican base. So if Paul voters are more likely to turn up in Obama precincts they are likely competing against fewer Republican voters than in a more Republican precinct. Hence the vote % for Paul will decline as total vote rises.

NH

In NH both Paul and Huntsman are hurt relative to Romney as total vote rises.

Looking at exit polling Huntsman's main demographic was Indies and Dems. With almost no Repub support. So it wasn't surprising when we found his % fell dramtically as you moved into less favorable areas for Obama.

Huntsman made the impact on Paul from Obama less (although still there). However, there was also a probably more important link made between Libertarian turnout vs Republican turnout. Using 2008 General election results, we found libertarian turnout lags in the areas where Republican turnout rises based on 2012 primary results. Since Paul is highly correlated with the libertarian demographic (areas where this demographic is stronger enables Paul to do better, the Libertarian party vote is a proxy for this) it would make sense that Paul's percentages would also lag as Republican turnout rises.

In combination, both the Huntsman and Paul numbers can be explained by Demographic features.

Other States

I believe these same type of analysis should explain the other states.

Other Candidates

In SC in 2008 it has been stated that Huckabee was also hurt by McCain in the same was Paul was hurt by Romney in many states in 2012 (these same results were also found in 2008 for Paul). While I haven't looked into this in depth my theory is that Huckabee caters to lower income voters that are more prominent in Obama areas than the wealthy McCain establishment voters. Hence an Obama correlation may exist, but for different reasons than Paul.

Conclusion

Demographic reasons such as Indie and Dem turnout relative to Republican turnout explain Huntsman's and Paul's declining cumulative vote percentage as a function of total precinct votes, which are correlated strongly with Republican turnout. Libertarian (not party, but belief system) concentration and turnout is also responsible for Paul's declining cumulative vote percentage. As the libertarian turnout falls relative to Republican turnout it is expected that Paul's percentage would also fall.

Based on the above I believe there is likely no total vote turnout fraud. Other fraud may exist, but the potential total vote turnout fraud was unusual for its dramatic impact on final results.

Appendix

Part I (VA)
http://www.ronpaulforums.com/showth...orithmic-vote-flipping.&p=4269744#post4269744

Part II (NH)
http://www.ronpaulforums.com/showth...orithmic-vote-flipping.&p=4272227#post4272227

 

[da32130]

FAQ

1) Are the libertarian votes in your analysis for NH (part 2) and for VA (part 3) too small to matter?

This is incorrect.

In NH there were 2748 votes over 305 precincts. The high was 75 and the low was 0. That is a range of 75 votes. Plenty for analysis. At best some of the smaller precincts are noisy, but the largers ones give a clear trend.

In VBC there were 1029 votes over 96 areas. The most was 22. And the two lowest were 0 and 2. With the rest in between. Plenty for analysis.

Just because a vote count in an area is 0,1,2,3,4,5 doesn't mean that isn't information. If the range is from 0-65 (NH), or 0-22 (VBC), falling anywhere along that range is informative. Because the best areas give 22 (or whatever the max percent is).

If you had a wealth cut off (1 million, 5 million, 10 million, or whatever works) that produced the same distribution I have with libertarians, it would give clues as to which areas wealthy people lived in.

The fact is the data works and performs exactly as expected.

2) Why not just use Barr?

In NH, Barr and Phillies are related. They are both libertarian.

In VA, Barr and Baldwin are related. Barr is a libertarian and Paul endorsed Baldwin.

They also add more data to the analysis giving us greater confidence in the results. However, the results are similar if we used just Barr.

 

[da32130]

Analysis of Alabama, delegate fraud. Could be just voters note voting for correct delegates. So Candidates revert to mean. (all move closer to the average delegate vote total)

****************************
Tuscaloosa - U of A Student Rec
*************************

Gingrich
Range 38-30
Candidate Total 30

Paul
Range 29-24
Candidate Total 26

Romney
Range 48-32
Candidate Total 53

Santorum
Range 35-30
Candidate Total 34

****************************
Tuscaloosa - Green Acres Health
*************************

Gingrich
Range 4-3
Candidate Total 1

Paul
Range 5-4
Candidate Total 4

Romney
Range 4-3 (a single 4, in the middle)
Candidate Total 3

Santorum
Range 9-8
Candidate Total 10

*******************
********************
********************


I was on the "wow, Alabama Fraud" side 2 days ago, but yesterday I did a bunch of research.

Anyway.

Some things to consider.

1) The "idiot effect" will effect candidates with fewer votes more than candidates with more votes.

2) The "idiot effect" will effect candidate at the top of the ballot more than the candidate at the bottom. From most to least, that's Gingrich, Paul, Romney, Santorum.

3) The "idiot effect" will effect least the candidate with the most idiots voting for them.

You will have a hard time finding instances where Santorum has more delegate votes, than votes for him. Not saying that Santorums voters are the idiots, of course.

I don't want to pick these numbers apart in great detail. I think that there are slightly more "idiots" than we think, skewing the numbers. 12% vote for all instead of 10% maybe. And the number of people voting only for who they're supposed to is a little lower.

 

[da32130]

This should apply to this thread as well. With the threads below giving further analysis about the above (along with the pro flipping threads). We would also ask the pro flippers to direct anti flippers here. Thank you.

Some negative derailing posts were deleted earlier from this thread. If anyone has criticisms of this project, please post them here:
http://www.ronpaulforums.com/showthread.php?367223-The-Case-Against-Vote-Flipping-(no-fraud)
http://www.ronpaulforums.com/showth...-PROOF-of-Vote-Flipping-Please-Prove-Me-Wrong.

This was the solution suggested by one of the administrators to the heated arguments and disruptions that were occurring in earlier threads about this topic, so this thread is to be kept free of that. It is a practice that has been adopted in the past to avoid these type of problems. I will update the OP with the link to the other threads.

 

[da32130]

FAQ

3. Do you think the pro flipping downwardly sloped graphs are an anomaly and evidence of fraud?

My point is this "the graphs aren't anomalies".

The anomaly would only be if they were done with a demographic analysis and there was a disparity.

I believe my graphs show the disparity can be explained by demographics (available - adjustments makes graph flat, and unavailable - adjustment should make graph flat).

When the other side presents an unadjusted sloped graph it hasn't even brought evidence to the table one way or another.

 

[da32130]

FAQ

4) Are smaller precincts really small in a population sense?

Not necessarily. They could be large and full of Democrats.

If republicans are few it means if indies and dems favor a candidate that candidate is likely to do very well in small precincts by total vote in a republican primary.

This is what happens to Paul in areas where indies and libertarians are stronger than normal vs republicans.

The reverse of this is probably what happens in Romney strongholds of high repub turnout vs Paul and Newt, etc. In other words, the far right of the graph where people suspect vote flipping, Romney has his base and should be expected to perform strongly.

Just like on the far left of the graph Paul does strongly. Which is explained by Indies and Libertarians.

We just don't have the same data available to prove this decisively for Romney. Then again, others can't prove we are wrong either.

 

[affa]

This 'debunk' has been debunked in about 5 ways so far. It's clear da32130 has no interest in addressing the many flaws in his understanding of the fraud argument, nor the many flaws in his methodology, including but not limited to 'weighting' charts based on statistically insignificant factors, ignoring that his 'proof' actually makes a case for fraud, false correlation, etc.

Not sure how I missed this thread. Probably because I assumed the 3 massive nasty existing threads were enough. Suppose not..

*Disclaimer* I don't think vote fraud explains this phenomenon. It's my realist bias. Thus far, however, nothing fully explains the precinct-size correlation.

da32130:

While it's good you are trying to incorporate demographics (+rep), you're making several mistakes.

1 -- this research was already done 15 days ago, with the data and analysis posted for everyone's use:

Revisited the data. This time, rolled precinct up and grouped by just the demographics and fields I wanted to test.

Fields Tested:
Median Income (County)
% White (County)
% Over 65 (County)
Precinct Size
Precinct % of County Vote
% Female (County)
People per Square Mile (County)
Precinct Turnout %

With all parameters, R^2 = .437 (explains roughly 44% of the noise)
With just the top 4, R^2 = .398 (explains 40% of the noise)
With just the top 3, R^2 = .256 (explains 26% of the noise) [--> this one excludes precinct size]

Median Income, % White, and Precinct Size have comparable significance (t ~ 15)
% Over 65 is slightly less significant (t = 9)

Romney's vote is positively correlated with Median Income, Precinct Size, and % Over 65. Interestingly, he's negatively correlated with % White.

For this type of rough analysis, an R^2 of .40 is significant. That .14 comes from Precinct Size is also significant.

From this analysis, precinct size cannot be ruled out as a significant independent variable. This suggests that something else must be driving the correlation (other demographic, campaign activity, fraud, etc).

2 -- You're assuming that if B and C are correlated with A, then B and C are correlated. That can't be assumed, it must be shown. Your analysis, as I mentioned in my PM and in the other thread, does not demonstrate that these factors are correlated with precinct size and therefore explain the phenomenon. You've shown that these factors are positively correlated with Romney vote %, but you haven't shown that they are positively correlated with precinct size! Nor, more specifically, that they fully explain the precinct-size correlation.

3 -- You're disregarding the conclusions of the analysis that was done correctly and did incorporate demographics. Precinct Size cannot be ruled out based on the demographics tested -- it's still significant in an Income/% White/% Over 65/Precinct Size multiple regression on Romney Vote %.

4 -- Because your analysis is incomplete and does not build upon prior analysis, your conclusion is misleading and hurts the efforts to understand the phenomenon. Truth is, we still can't fully explain what's driving the precinct size / Romney positive correlation

5 -- You should release your data and analysis for peer review

6 -- Your final weighted graph assumes each of your variables are independent and not correlated. However, female % and republicanism are anti-correlated. They essentially measure the same thing, and are therefore double counted. The pivot table graph in the file below lets you view this type of information easily. You'll find that there is still precinct-size correlation even when the other demographics are normalized, just as the analysis in (1) above shows.

Still, most shocking of all, is that no one else has pointed out the obvious errors with your approach (#2, #6). Then again, based on the "analysts" in the other thread, I suppose it shouldn't be too surprising.

For the 50th time, if someone wants to do some demographic analysis, do not start from scratch!! Use this file:

http://dl.dropbox.com/u/63868969/RPF/SC_2012_Primary_Analysis.xlsx



Conclusion: Not debunked

Here's another way to look at the Virginia Beach City graph from da32130 above. The graph below is the same as da32130's graph #1 but zoomed in a bit. In the segment that includes the 25 precincts where da32130 could not explain Romney's gains by demographics, the R- squared value between Romney's curve and its best- fit line is .9837. In layman's terms, this means that the variation in Romney's cumulative percentage is 99.2% explained simply by the number of votes cast per precinct.. forget demographics. Put that together with my previous post, and you get Electronic Vote Manipulation... PERIOD.
BTW- when da32130 speaks of explaining the gains in "85- 90% of the precincts," he fails to point out that the remaining 10- 15% that he cannot account for represent 40% of the total vote! Hence, this is why the riggers choose these larger precincts to manipulate: more votes to gain and harder to get caught.

And finally:

The Man, if I only had Obama%, the charts would look even more like flipping. However, my argument would be that Obama% explained where I expected. So it is likely that variables such as libertarian% and republican demographic data could explain the rest if they existed.
SNIP
Fortunately, the Libertarian party exists so we can make that adjustment. Otherwise, how could I convince you that Paul would do better in libertarian areas? That would be even tougher to prove.

I have asked you repeatedly to stop using and weighting your charts by 'libertarian' results, because they are not only statistically insignificant, but highly volatile across election cycles. This is an extremely basic concept, which I have spelled out to you again and again. Yet you insist on doing it over and over, creating straight up deceitful charts by 'weighting' your own charts based on data that you have been warned is not reliable, consistent, or significant.

But fine. I'll prove it to you. In fact, I'll go one further and show you why not only is using Libertarian performance not reliable, consistent, or significant, but even using Democrat/Republican performance from 2008 as a indicator for 2012 is dangerous due to issues such as redistricting, Obamamania, and Bush fatigue, all of which make 2008 results potentially unrepresentative of normal voting habits.

If you continue to weight your charts by '2008 Libertarian Performance', or continue to fail to take into account the fallibility of 2008 results as demographic indicators of a precinct, I have no choice but to come to the conclusion you are intentionally obfuscating the facts.

Libertarian votes from 2008 can not, and should not, be used to weight 2012 results, because they are statistically insignificant, unreliable, and inconsistent across elections. It's dangerous even using 2008 Democratic and Republican results as a solid indicator of voting habits per precinct, given the massive differences seen in some precincts from 2004 to 2008.

If you want to study demographics, all I can suggest is using current demographic data.

 

[da32130]

This 'debunk' has been debunked in about 5 ways so far.

Affa, I thought you were suppose to stay off of this thread?

These points have been answered. You just don't like those answers.

I will respond to them one by one for others using my prior answers. Thank you for compiling them for me.

 

[da32130]

I guess this should go both ways since it was posted on a pro flipper thread:

If you have nothing positive to add to this discussion, do not post. Stop derailing the thread.

 

[affa]

Affa, I thought you were suppose to stay off of this thread?

These points have been answered. You just don't like those answers.

I will respond to them one by one for others using my prior answers. Thank you for compiling them for me.

Go ahead, continue to 'weight' your charts based on data hand picked by yourself to show what you want, even though the numbers you're choosing to 'weight' by have been proven to be statistically insignificant, unreliable, and highly volatile from election to election. Have fun with your thread, sir.

 

[soulcyon]

What is this, I don't even...

da32130 - these points have been debunked over and over again. If you're going to make an honest debunk, at least have the courtesy to properly address what you are debunking. You seem to be slashing and burning at generic points that you cooked up yourself and not a claim anybody has made specifically.

 

[sailingaway]

I guess this should go both ways since it was posted on a pro flipper thread:

parallel attacks in a different thread are still attacks, and I will move this thread if they continue. The other thread isn't mocking you.

 

[da32130]

//

 

[da32130]

parallel attacks in a different thread are still attacks, and I will move this thread if they continue. The other thread isn't mocking you.

Sailing, i was being serious. The threads shouldn't spam each other.

If affa continues to post things i've already answered. Isn't that spam? Isn't that derailing the thread?

He is just using fancier words to mock and distort.

I've literally answered most of his points in my OP and FAQs. Not to mention in my prior threads.

 

[da32130]

What is this, I don't even...

da32130 - these points have been debunked over and over again. If you're going to make an honest debunk, at least have the courtesy to properly address what you are debunking. You seem to be slashing and burning at generic points that you cooked up yourself and not a claim anybody has made specifically.

They have not been debunked. They have been ignored or we have agreed to disagree. That is why I am reposting. So people know that.

However, you can't expect the anti flippers to expel the same energy as the pro flippers. By definition the anti-flippers would like to move on to more important things (to them).